Light-field cameras, which may also be referred to as plenoptic cameras, use a plenoptic microlens array (MLA), in combination with a photosensor array, to capture directional information of light rays passing through the camera's optics. Such directional information can be used for providing and implementing advanced display of and interaction with captured pictures, such as refocusing after capture. Such techniques are described, for example, in Ng et al., “Light Field Photography with a Hand-Held Plenoptic Camera”, Technical Report CSTR 2005-02, Stanford Computer Science, and in related U.S. Utility application Ser. No. 12/632,979, for “Light-field Data Acquisition Devices, and Methods of Using and Manufacturing Same,” filed Dec. 8, 2009, the disclosure of which is incorporated herein by reference.
Plenoptic microlens arrays are often manufactured using a polymer-on-glass approach, including a stamping or replication process wherein the plenoptic MLA is fabricated as a polymer attached to a transparent glass surface. Plenoptic MLAs can be constructed in such a manner using machines and processes available, for example, from Suss MicroOptics of Neuchatel, Switzerland. The polymer-on-glass MLA array is placed with the lens side down, such that incoming light passes through the glass and is then directed by the plenoptic MLA onto the surface of a photosensor array.
Referring now to FIG. 1A, there is shown an example of an assembly 100 for a light-field camera according to the prior art, wherein the plenoptic MLA 102, including any number of individual microlenses 116, is constructed using a polymer-on-glass approach, resulting in MLA 102 being fabricated on glass 103. An air gap 105 has been introduced between plenoptic MLA 102 and photosensor array 101 of the device, to allow for light rays to be properly directed to correct locations on photosensor array 101.
In general, existing techniques for manufacturing a light field sensor require that photosensor array 101 and plenoptic MLA 102 be fabricated as separate components. These components may be assembled using a mechanical separator that adds air gap 105 between the components. Such an assembly process can be expensive and cumbersome; furthermore, the resulting air gap 105 is a potential source of misalignment, unreliability, and/or reduced optical performance. It is desirable to avoid such separate fabrication of parts and later assembly using mechanical separation so as to improve manufacturing efficiency, and so that precision in placement of the lens components can be achieved.
In many image capture devices, a different type of microlens array, referred to herein as a pixel-level microlens array, is used to improve light capture performance and/or reduce crosstalk between neighboring pixels in a photosensor array 101. Referring now to FIG. 1B, there is shown an example of an assembly 150 according to the prior art. Relative to FIG. 1A, this diagram is shown at much higher magnification. Microlenses 206 in pixel-level microlens array 202 direct incoming light 104 so as to maximize the amount of light that reaches each individual photosensor 106, and to avoid losing light that would otherwise hit the areas between individual photosensors 106. Such an arrangement is well known in the art, and may be included on many commercially available image sensors.
The plenoptic microlens array 102 depicted in FIG. 1A and the pixel-level microlens array 202 depicted in FIG. 1B serve completely different purposes. In general, these two types of microlens arrays are constructed to be of differing sizes and locations. For example, each microlens 206 of pixel-level microlens array 202 may be approximately 2 microns across, while each microlens 116 of the plenoptic microlens array 102 may be approximately 20 microns across. These measurements are merely examples. In general, pixel-level microlenses 206 may have a 1:1 relationship with photosensors 106, while plenoptic microlenses 116 may have a 1:many relationship with photosensors 106.
Referring now to FIG. 2, there is shown an optical assembly 200 including both a plenoptic MLA 102 and a pixel-level MLA 202 according to the prior art. Such an assembly 200 effectively combines the components depicted and described in connection with FIGS. 1A and 1B. Here, plenoptic MLA 102 directs incoming light 104 toward pixel-level MLA 202. Microlenses 206 in pixel-level MLA 202 then further direct light toward individual photosensors 106 in photosensor array 101. In the arrangement of FIG. 2, air gap 105 is provided between plenoptic MLA 102 and pixel-level MLA 202.
As described above, plenoptic MLA 102 of FIG. 2 can be constructed using a polymer-on-glass approach, wherein plenoptic MLA 102 is attached to glass surface 103. For example, plenoptic MLA 102 may be formed using a mold that is stamped out using polymer and affixed to glass surface 103. The resulting plano-convex microlens assembly is positioned in a “face-down” manner as shown in FIG. 2, with the convex lens surfaces facing away from the light source.
The inclusion of both a plenoptic MLA 102 and a pixel-level MLA 202 serves to further complicate the construction of the image capture apparatus. Existing techniques offer no reliable method for constructing an image capture apparatus that employs both a plenoptic MLA 102 and a pixel-level MLA 202, without introducing an air gap 105. Introduction of such an air gap 105 potentially introduces further complexity, cost, and potential for misalignment.